A soil heavy metal pollution identification system

Abstract

The application relates to the technical field of soil pollution identification, and discloses a soil heavy metal pollution identification system. A multispectral remote sensing perception module of the system obtains surface reflectivity data and soil in-situ spectrum data through satellite load and a vehicle-mounted mobile platform respectively, and generates a multi-band spectrum response signal through a heterogeneous data fusion gateway. In a pollution risk assessment module, a spatial distribution analysis unit outputs a heavy metal spatial distribution atlas, a migration risk prediction unit generates a pollution migration probability cloud map in combination with meteorological and hydrological data, and a pollution threshold definition unit outputs a soil remediation safety threshold. In a governance decision execution module, an in-situ remediation execution unit adjusts passivation agent injection parameters, a pollution source control unit controls pollution source blocking equipment and collects monitoring signals, and a three-dimensional dynamic early warning platform generates a comprehensive pollution risk index. The system realizes the collaborative operation of soil heavy metal pollution identification, evaluation and treatment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of soil pollution identification, in particular to a soil heavy metal pollution identification system. BACKGROUND

[0002] There are many limitations in the identification and evaluation of soil heavy metal pollution. The traditional soil sampling analysis method relies on manual collection of soil samples, and then obtains pollution data through laboratory detection. This method not only consumes time and effort, but also has limited sampling points, making it difficult to fully reflect the pollution status of large-area soil and easily missing local areas with serious pollution.

[0003] In the aspect of remote sensing monitoring, the existing technology mostly uses single satellite remote sensing data, which can achieve wide range monitoring, but is limited by spatial resolution and spectral resolution, making it difficult to accurately identify the types and contents of heavy metals in soil. At the same time, the combination of ground spectral collection and satellite remote sensing data is not high, and the data of the two are often independent of each other, which cannot form effective complementation, resulting in low utilization efficiency of spectral information.

[0004] In the pollution risk assessment link, most methods can only simply describe the spatial distribution of heavy metals, lack in-depth analysis of their migration rules, and are difficult to predict the trend of pollution diffusion. In addition, the definition of pollution threshold is mostly based on empirical values or single factors, and cannot be dynamically adjusted combined with multi-source spectral data, affecting the accuracy and applicability of the threshold.

[0005] In the aspect of treatment decision, the repair execution and pollution source control of the existing system are often independent of each other, lacking a collaborative operation mechanism. The adjustment of repair measures cannot timely refer to real-time data of pollution migration, and the pollution source control is also difficult to accurately control according to the actual threshold of soil remediation, resulting in unsatisfactory treatment effect and difficulty in forming a closed pollution treatment system. SUMMARY

[0006] The purpose of the present application is to provide a soil heavy metal pollution identification system to solve the problems raised in the background.

[0007] To achieve the above purpose, the present application provides a soil heavy metal pollution identification system, which comprises:

[0008] A multispectral remote sensing perception module is configured with a high-resolution imaging unit and a ground spectral collection unit. The high-resolution imaging unit obtains ground reflectance data through satellite payload, and the ground spectral collection unit collects in-situ soil spectral data using a vehicle-mounted mobile platform. The multispectral remote sensing perception module is integrated with a heterogeneous data fusion gateway, which receives ground reflectance data and in-situ soil spectral data and generates a multi-band spectral response signal.

[0009] The pollution risk assessment module includes a spatial distribution analysis unit, a migration risk prediction unit, and a pollution threshold definition unit. The spatial distribution analysis unit processes multi-band spectral response signals and outputs heavy metal spatial distribution maps. The migration risk prediction unit combines heavy metal spatial distribution maps and meteorological and hydrological data to generate pollution migration probability cloud maps. The pollution threshold definition unit defines ecological safety constraints based on multi-band spectral response signals and outputs soil remediation safety thresholds.

[0010] The governance decision execution module includes an in-situ remediation execution unit, a pollution source control unit, and a three-dimensional dynamic early warning platform. The in-situ remediation execution unit receives pollution migration probability cloud maps and adjusts phasing agent injection parameters. The pollution source control unit regulates the operating state of pollution source blocking equipment based on soil remediation safety thresholds and multi-band spectral response signals, while collecting real-time monitoring signals of soil pore water. The three-dimensional dynamic early warning platform integrates heavy metal spatial distribution maps and pollution migration probability cloud maps to generate a comprehensive pollution risk index.

[0011] Preferably, the multi-spectral remote sensing perception module further includes a UAV near-ground detection unit and an underground water quality sensing network. The UAV near-ground detection unit obtains vegetation stress characteristic data by carrying a hyperspectral sensor on a multi-rotor flight platform. The underground water quality sensing network is deployed in monitoring wells at pollution sites and collects underground water ion concentration data. The heterogeneous data fusion gateway receives vegetation stress characteristic data and underground water ion concentration data, performs spatio-temporal alignment processing, and forms multi-band spectral response signals containing spectral feature signals, vegetation response signals, and ion diffusion signals. The multi-band spectral response signals are specifically divided into spatial analysis spectral signals, migration prediction spectral signals, and threshold definition spectral signals.

[0012] Preferably, the spatial distribution analysis unit uses a convolutional neural network to construct a geostatistical inversion model. The spatial analysis spectral signals include visible-near-infrared band reflectance data, short-wave infrared absorption feature data, and historical pollution census data. The geostatistical inversion model extracts spectral spatial features through a depth separable convolutional layer, inputs visible-near-infrared band reflectance data, short-wave infrared absorption feature data, and historical pollution census data into a three-dimensional convolution kernel to generate heavy metal spatial distribution maps, and outputs pollution source tracing results by fusing vegetation response signals and ion diffusion signals through a feature pyramid network.

[0013] Preferably, the migration risk prediction unit comprises a pollution migration model and a Monte Carlo simulator, the migration prediction spectrum signal comprises soil texture parameters, groundwater flow velocity data and rainfall erosion intensity data, the pollution migration model processes the migration prediction spectrum signal by coupling a convection-diffusion equation with an adsorption-desorption kinetics equation, outputs a heavy metal migration flux prediction value, the Monte Carlo simulator generates a pollution migration probability cloud based on the heavy metal migration flux prediction value through a thousand times of random sampling operation, and the pollution migration probability cloud is fed back to the geo-statistical inversion model to update the pollution source tracing result.

[0014] Preferably, the pollution threshold defining unit processes the threshold defining spectrum signal through an ecological toxicology constraint model, the ecological toxicology constraint model is integrated with a species sensitivity distribution analysis module, the species sensitivity distribution analysis module calls biological half lethal concentration data in a local biological database to define an ecological safety constraint condition, outputs a soil remediation safety threshold, and simultaneously adopts a model predictive control algorithm to dynamically adjust the soil remediation safety threshold to be not lower than a legal limit standard.

[0015] Preferably, the in-situ remediation execution unit is configured with a pollution migration model and a passivation agent optimization library, the pollution migration model is constructed based on Fick's second law, a three-dimensional diffusion equation, input parameters of the three-dimensional diffusion equation include soil oxidation-reduction potential data, organic matter content data and the pollution migration probability cloud, and outputs an optimal passivation agent dosage signal, the passivation agent optimization library stores spectral response characteristic data of historical remediation materials, and a passivation agent type selection instruction is generated by matching the spectral response characteristic data with the optimal passivation agent dosage signal.

[0016] Preferably, the pollution source control unit comprises a pollution flux monitoring sub-module and a blocking equipment regulation sub-module, the pollution flux monitoring sub-module receives real-time monitoring signals of soil pore water and heavy metal spatial distribution maps, calculates heavy metal leakage flux in a pollution source area, and the blocking equipment regulation sub-module generates a vertical barrier wall depth adjustment instruction based on a difference between the heavy metal leakage flux and the soil remediation safety threshold, the vertical barrier wall depth adjustment instruction is transmitted to the three-dimensional diffusion equation to update the optimal passivation agent dosage signal.

[0017] Preferably, the three-dimensional dynamic early warning platform is deployed with a pollution heat map generation engine and a remediation scheme deduction module, the pollution heat map generation engine renders the heavy metal spatial distribution map into a three-dimensional pollution heat map through a Kriging interpolation algorithm, and the remediation scheme deduction module generates a multi-scheme remediation path based on a comprehensive pollution risk index calling a historical remediation case library, the multi-scheme remediation path is input into the in-situ remediation execution unit to trigger dynamic adjustment of passivation agent parameters.

[0018] Preferably, the three-dimensional dynamic early warning platform implements a multi-level risk response mechanism, when the comprehensive pollution risk index reaches the primary warning threshold, the pollution source area deceleration production instruction is started, when the comprehensive pollution risk index reaches the intermediate warning threshold, the pollution site isolation protection instruction is activated and the in-situ remediation execution unit is triggered to urgently add the passivation agent, when the comprehensive pollution risk index reaches the highest warning threshold, the site emergency closure signal is sent to the governance decision execution module.

[0019] Preferably, the in-situ remediation execution unit specifically comprises an intelligent grouting system and a passivation agent reaction monitoring device, the intelligent grouting system analyzes the optimal passivation agent dosage signal and dynamically adjusts the grouting pressure and grouting flow, the passivation agent reaction monitoring device detects the heavy metal chelation rate of the passivation agent reaction product in real time through a laser-induced breakdown spectroscopy technology, generates a passivation agent performance feedback signal, and the passivation agent performance feedback signal is transmitted to the pollution migration model to update the three-dimensional diffusion equation parameters.

[0020] Compared with the prior art, the beneficial effects of the present application are:

[0021] The system realizes the organic fusion of satellite remote sensing and ground spectrum collection through a multispectral remote sensing perception module. The surface reflectance data obtained by the high-resolution imaging unit and the soil in-situ spectrum data collected by the vehicle-mounted mobile platform generate a multi-band spectral response signal after being processed by a heterogeneous data fusion gateway, integrating the advantages of different sources of spectral data, and making the spectral information of soil heavy metal pollution more comprehensive and rich.

[0022] In the pollution risk assessment module, the processing of the multi-band spectral response signal by the spatial distribution analysis unit can present a clear heavy metal spatial distribution map, and intuitively display the spatial pattern of pollution. The pollution migration probability cloud map generated by the migration risk prediction unit in combination with meteorological and hydrological data can clearly reflect the migration trend of heavy metals in the soil. The pollution threshold definition unit defines the ecological safety constraint conditions according to the multi-band spectral response signal, so that the output soil remediation safety threshold is more in line with the actual soil environment conditions, and can better adapt to the pollution characteristics of different regions.

[0023] The governance decision execution module combines in-situ remediation, pollution source control and dynamic early warning to form a complete governance chain. The in-situ remediation execution unit adjusts the passivation agent injection parameters according to the pollution migration probability cloud map, so that the remediation measures can flexibly change with the pollution migration situation. The pollution source control unit regulates the equipment operating state based on the soil remediation safety threshold and the multi-band spectral response signal, and simultaneously collects the soil pore water monitoring signal, so that the pollution source blocking is more targeted. The three-dimensional dynamic early warning platform integrates multiple types of graphs to generate a comprehensive pollution risk index, which can comprehensively present the pollution risk situation and facilitate timely grasp of the overall pollution situation.

[0024] Data transmission between modules is smooth, forming a complete process from data perception, risk assessment to decision execution, realizing integrated operation of soil heavy metal pollution identification, assessment and treatment, and improving the comprehensive management and control ability of soil heavy metal pollution. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 The working principle diagram of the soil heavy metal pollution identification system is provided.

[0026] Figure 2 The working flow chart of the multispectral remote sensing perception module is provided.

[0027] Figure 3 The working flow chart of the migration risk prediction unit is provided.

[0028] Figure 4 The working flow chart of the pollution source management and control unit is provided. DETAILED DESCRIPTION

[0029] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor are within the scope of protection of the present application.

[0030] Please refer to Figure 1 The present application provides a soil heavy metal pollution identification system, which comprises:

[0031] Precise pollution identification and repair are realized through closed-loop control of remote sensing perception, risk assessment and treatment execution. The system hardware is composed of satellite payload, vehicle-mounted mobile platform, underground sensor network and repair equipment, and the software algorithm includes heterogeneous data fusion engine, deep learning model and dynamic early warning mechanism. When the system is running, the high-resolution imaging unit obtains the ground reflectivity data of 400-2500nm wave band through the satellite multispectral sensor, the ground spectral acquisition unit adopts the vehicle-mounted ASDFieldSpec spectrometer to collect the in-situ soil spectrum with 1nm resolution, and the two are radiometrically corrected and geometrically registered through the heterogeneous data fusion gateway to generate multispectral response signals containing visible light, near-infrared and short-wave infrared features.

[0032] Embodiment 1: refer to Figure 2, involving the detailed construction and operation mechanism of the multispectral remote sensing perception module, focusing on the cooperative workflow of the unmanned aerial vehicle near-ground detection unit, the underground water quality sensor network, and the heterogeneous data fusion gateway. The system hardware configuration uses the DJI M300RTK six-rotor unmanned aerial vehicle platform as the air carrier, which is equipped with a dual-frequency GNSS receiver and an RTK positioning module, capable of achieving centimeter-level positioning accuracy. The flight control system presets a grid flight route, and the flight height is stably maintained at 100 meters above ground, with a flight speed controlled at 3 meters per second, ensuring stable data acquisition conditions for the hyperspectral sensor. The HySpex VNIR-1800 imaging spectrometer carried by the unmanned aerial vehicle has a working wavelength range of 400-1000 nm, a spectral resolution of 3.5 nm, and a spatial resolution of 5 cm. The sensor has a built-in three-axis stabilized gimbal that automatically compensates for image shifts caused by attitude changes during flight.

[0033] The groundwater quality monitoring network consists of 12 monitoring wells distributed around the contaminated site. The well depth is determined according to the groundwater flow direction, with 3 wells arranged upstream, 6 wells arranged downstream, and 3 wells arranged laterally. Each monitoring well is equipped with a YSI EX02 multi-parameter water quality sensor, and the sensor probe contains six measurement channels for detecting pH, dissolved oxygen, conductivity, temperature, oxidation-reduction potential, and turbidity. The concentration of heavy metal ions is detected using an independent voltammetric analysis module, which measures the content levels of elements such as Cd, Pb, and As through anodic stripping voltammetry. Multiple sampling ports are set up in the monitoring well, with sensor nodes arranged at depths of 5 meters, 10 meters, and 15 meters, forming a three-dimensional monitoring network. All sensor nodes are connected to the wellhead data collector through an RS-485 bus, and the collector has a built-in 4G communication module for remote data transmission.

[0034] The heterogeneous data fusion gateway is deployed on an edge computing server, with hardware configuration using an Intel Xeon E5 processor and 64 GB of memory, running the Ubuntu operating system. The data receiving port has four independent channels: the satellite data channel receives the Landsat 9 surface reflectance product updated at 12 o'clock every day through the FTP protocol; the vehicle-mounted spectral data channel uses the TCP / IP protocol to transmit the measurement results of the ASD FieldSpec spectrometer in real time; the unmanned aerial vehicle data channel receives the hyperspectral cube data returned after the flight task is completed through the WiFi6 wireless network; and the groundwater data channel obtains the real-time monitoring values of the sensor network through the MQTT protocol. The spatiotemporal alignment process uses a PostgreSQL database to store all raw data and establishes a spatiotemporal index table to record the collection time and spatial coordinates of each data point.

[0035] In the data preprocessing stage, satellite image data is first radiometrically calibrated and atmospherically corrected to convert digital quantization values into surface reflectance. Vehicle-mounted spectral data is smoothed to eliminate instrument noise, and reflectance calibration is performed using a standard whiteboard. UAV hyperspectral data is geometrically corrected to eliminate lens distortion and stitched to generate a continuous orthophoto. Groundwater data is quality controlled to remove outliers outside the sensor range. The preprocessed data is converted to the UTM coordinate system and the timestamp is converted to UTC time.

[0036] The spatio-temporal fusion algorithm uses a three-layer processing architecture. The first layer implements spatial registration, taking satellite images as the reference, and aligns other data sources to the same spatial reference system through feature point matching. The second layer performs time interpolation, and the cubic spline interpolation method is used to generate synchronous time series for data sources with different sampling frequencies. The third layer performs data fusion. For spectral data, principal component analysis is used to reduce dimensionality, and hyperspectral data and multispectral data are fused into a data cube containing 20 characteristic bands. For water quality data, Kriging spatial interpolation is performed to generate a continuously distributed groundwater quality parameter field.

[0037] The generation process of multi-band spectral response signals includes two stages of feature extraction and signal classification. The feature extraction module identifies three key features from the fused data cube: the vegetation green peak reflectance at 550 nm is extracted as the vegetation stress indicator; the clay mineral absorption depth at 2200 nm is extracted as the soil pollution indicator; and the iron oxide reflectance feature at 650 nm is extracted as the groundwater pollution diffusion marker. The signal classifier uses a random forest algorithm to divide the input data stream into three categories of output according to the feature importance score: spatial analysis spectral signals mainly contain surface reflectance features and soil property parameters; migration prediction spectral signals focus on groundwater quality changes and permeability coefficients; threshold-defined spectral signals integrate biological toxicity data and environmental standard limits.

[0038] The implementation of the deep separable convolutional network is based on the TensorFlow framework, and the network input layer is set to a three-dimensional tensor of 256x256x20, corresponding to the spatial distribution of 20 characteristic bands. The first convolutional layer uses 32 3x3x5 convolutional kernels for spatial feature extraction, and each convolutional kernel slides independently in three dimensions. The batch normalization layer normalizes each feature channel, and the ReLU activation function introduces non-linear transformation. Subsequent network layers gradually increase the number of convolutional kernels to 128, while reducing the spatial resolution through max-pooling layers. Short-wave infrared absorption feature data is fused with the main network through a jump connection at the fourth convolutional layer, and historical pollution survey data is converted to a 64-dimensional vector through an embedding layer and connected to a fully connected layer.

[0039] The feature pyramid network builds a top-down feature transmission path, and the high-resolution features extracted by the bottom network are gradually fused with the semantic features of the high-level network. The vegetation response signal processing branch uses a two-level cascaded LSTM network to analyze the NDVI time series change characteristics. The ion diffusion signal processing branch builds a graph convolution network, and the topological structure is established by taking the monitoring well as the node and the groundwater flow path as the edge. The final output of the heavy metal spatial distribution map is stored in GeoTIFF format, including five pollution level classification results and the spatial concentration distribution of three heavy metals. The pollution source tracing algorithm is based on the back propagation particle tracking model, combined with groundwater flow field data and pollutant diffusion coefficient to calculate the possible pollution source position, and mark the high probability pollution source area with a red polygon in the map.

[0040] Embodiment 2: see Figure 3 , which involves a collaborative computing system for pollution migration risk prediction and ecological safety threshold definition, focuses on the dynamic evaluation method combining physical process modeling and stochastic simulation. The migration risk prediction unit is deployed on a high-performance server equipped with NVIDIA Tesla V100 computing cards, running a numerical simulation environment based on multi-physical field coupling. Three types of basic data need to be loaded in the system initialization stage: soil texture parameters are measured by a Malvern laser particle size analyzer, the sample is screened and dispersed, and the measurement results are converted into the USDA soil classification standard; groundwater flow rate data come from a distributed optical fiber sensing system, and the flow rate distribution is obtained by inverting the temperature field change monitored by DTS through the heat pulse method; rainfall erosion intensity data come from the precipitation intensity observation of meteorological station millimeter wave radar, and the ground rainfall spatial distribution is obtained through Z-R relationship conversion.

[0041] The pollution migration model adopts a modular architecture design, including six main functional components. The solute transport component realizes the numerical solution of the convection-diffusion equation, and the unstructured tetrahedral grid is adopted for spatial discretization to adapt to complex geological conditions, and the Crank-Nicolson format is adopted for time advancement to ensure calculation stability. The adsorption and desorption component integrates multiple constitutive relations, including linear distribution model, Freundlich isotherm and Langmuir equation, and automatically selects the applicable model according to the type of pollutants. The groundwater flow component solves the modified Richards equation to handle the coupled flow in the unsaturated and saturated zones. The chemical reaction component includes basic reaction types such as oxidation-reduction, precipitation-dissolution and surface complexation, and describes the interaction of multiple components through a stoichiometric matrix. The parameter estimation component uses the adjoint state method to automatically calibrate model parameters, reducing errors caused by initial condition uncertainty. The visualization component generates dynamic simulation results to display the temporal and spatial evolution of pollutant concentration with color contour maps.

[0042] The Monte Carlo simulator is built on the R language platform, including three core modules: parameter perturbation generator, parallel task scheduler, and result statistical analyzer. The parameter perturbation generator uses the Latin hypercube sampling technique to sample 12 key input parameters, including porosity, permeability, and distribution coefficient. Each parameter has a reasonable physical range, and the number of sampling points is set to 1000 to ensure the probability coverage. The parallel task scheduler allocates computing resources through the SLURM job management system, and each sampling point is submitted to the computing cluster as an independent task. The result statistical analyzer collects all simulation results, calculates the statistical quantities of pollutant concentration at each spatial location, including mean, standard deviation, and percentile.

[0043] The generation process of the pollution migration probability cloud map includes four processing steps. The first step is to implement data spatialization, which interpolates the concentration field output by the simulation to a regular grid with a resolution of 10 meters to match the scale of remote sensing data. The second step is to perform probability calculation, which counts the number of simulations that exceed the specified threshold for each grid cell and converts it to a probability value. The third step is to implement smoothing processing, which uses an anisotropic Gaussian filter to eliminate numerical noise and maintain the spatial continuity of the main pollution plume. The fourth step is to perform coordinate conversion, which converts the calculation results from the local model coordinate system to the WGS84 geographic coordinate system, making it easier to overlay and display with other spatial data. The final cloud map uses a hierarchical coloring scheme, with warm colors representing high-risk areas and cold colors representing low-risk areas, and is displayed on the three-dimensional geographic information system base map in a semi-transparent overlay manner.

[0044] The architecture design of the ecological toxicity constraint model follows the standard process of species sensitivity distribution analysis. The local biological database stores strictly screened toxicity data, covering 112 species of fish, crustaceans, algae, and soil invertebrates. Quality control is implemented during the data preprocessing stage to remove data records with unclear experimental conditions or non-standard statistical methods. For multiple test data of the same pollutant, the arithmetic mean method is used for integration. The maximum likelihood estimation method is used to solve the parameters of the log-logistic distribution during model fitting, and the Bootstrap resampling technique is used to evaluate the parameter uncertainty. The hazard concentration calculation module implements four statistical extrapolation methods, including HC5 calculation, confidence interval estimation, model goodness-of-fit test, and distribution function visualization.

[0045] The running process of the species sensitivity distribution analysis module includes four stages: data import, model fitting, result verification, and report generation. The data import stage supports input files in CSV and Excel formats, automatically identifies concentration units and test species information. The model fitting stage provides five distribution function options, with the default being the log-logistic distribution. The result verification stage implements residual analysis and Q-Q plot testing to evaluate the goodness of fit of the model. The report generation stage outputs a PDF document containing fitted curve graphs, parameter estimation tables, and statistical test results. During the analysis process, a data integrity checkpoint is set, and when the input data is less than 10 groups, the system automatically switches to a safety threshold calculation mode.

[0046] The implementation of the model predictive control algorithm is based on the Python control library, which includes three components: a state observer, a dynamic model, and an optimization solver. The state observer monitors the calculation process of the soil remediation safety threshold in real time and records the historical trajectory of each parameter update. The dynamic model uses a first-order lag element to describe the inertia characteristics of threshold changes, and sets a reasonable time constant to reflect the response delay of the ecological system. The optimization solver constructs a constrained nonlinear programming problem, with the legal limit standard as a hard constraint condition, and iteratively solves the optimal threshold value by the interior point method. During the algorithm running process, a double-check mechanism is set, and when the calculated value approaches the limit standard, an artificial review process is automatically triggered.

[0047] The threshold dynamic adjustment process is based on a feedforward-feedback composite control strategy. The feedforward element predicts the threshold adjustment direction according to the pollution source intensity changes, and the feedback element eliminates system deviations through error integration. The control parameter tuning uses the Ziegler-Nichols method, and fine-tuning is obtained through trial and error to achieve stable control effect. The adjustment instruction generation module outputs standardized control signals, which are transmitted to the remediation equipment actuators through the OPCUA protocol. The system sets a threshold change rate limit to prevent large fluctuations in a short period of time from affecting the remediation effect. Historical threshold data is stored in a time series database, supporting time range queries and trend analysis.

[0048] The output format of the soil remediation safety threshold uses the JSON standard, including threshold value, effective time, applicable area, and review mark fields. The threshold application module implements spatial differentiation management, divides different risk areas according to the spatial distribution map of heavy metals, and sets graded thresholds respectively. When the monitoring data exceeds the current threshold, the system automatically triggers an early warning signal and records the violation event. The threshold review mechanism sets a regular evaluation period, recalculates the safety threshold based on the latest monitoring data, and maintains the timeliness of risk control measures. All threshold change operations are recorded on the blockchain evidence platform to ensure data integrity and traceability.

[0049] Example 3: see Figure 4, the collaborative working mechanism of the in-situ remediation execution unit and the pollution source management unit is carried out, and the technical implementation path of the optimization of the passivator dosage and the regulation of the vertical barrier wall is emphasized. The three-dimensional diffusion equation of the in-situ remediation execution unit is based on the solute transport theory in the heterogeneous porous medium, and the finite volume method is used for spatial discretization processing. The soil oxidation-reduction potential monitoring network is composed of 48 platinum electrodes, which are arranged according to a 20 meter by 20 meter grid, with three depth layers (0.5 meters, 1.5 meters and 3.0 meters) at each measurement point. The measurement value is recorded by a multi-channel data acquisition instrument every 15 minutes. The organic matter content determination adopts the potassium dichromate oxidation method, and the sample pretreatment process includes grinding and sieving and hydrochloric acid removal of inorganic carbon. The measurement results are generated by an interpolation algorithm to generate a spatial distribution field.

[0050] The construction process of the passivator optimization library includes material characterization, spectrum filing and performance evaluation. 23 kinds of repair materials are analyzed by X-ray diffraction and Fourier transform infrared spectroscopy to establish a database of crystal structure and functional group characteristics. Each material sets up an independent spectrum response characteristic record, including the reflectivity curve of 400-2500 nm waveband, the characteristic absorption peak position and the relative intensity parameter. The material performance data come from standard batch experiments, recording the immobilization efficiency of heavy metals such as Cd, Pb and As under different environmental conditions. The database uses a hierarchical storage structure, with the top layer classified by material type, the middle layer indexed by pollutant species, and the bottom layer storing specific experimental condition and treatment effect data sets.

[0051] The control logic of the intelligent grouting system is based on the concentration gradient field output by the diffusion equation, which divides the repair area into multiple control units. The grouting pressure regulation adopts a hierarchical control strategy, with the basic pressure set to 0.3 MPa, and the pollution level increased by 0.1 MPa in a stepwise manner, with the maximum working pressure limited to 1.2 MPa. The flow control system uses an electromagnetic flowmeter for real-time monitoring, and adjusts the frequency of the pump through a PID algorithm to maintain the set flow value within ±5% error range. The grouting point layout follows the spatial distribution characteristics of pollutants, with dense grouting holes preferentially set in high concentration areas, and the hole spacing dynamically adjusted according to the pollution plume width, with the minimum spacing set to 5 meters.

[0052] The heavy metal leakage flux calculation model of the pollution source management unit integrates Darcy's law and the solute flux equation:

[0053]

[0054] where represents the total leakage flux (kg / d), is the soil dry density (g / cm³) of the i-th monitoring point, is the pore water velocity (m / d), represents the control section area (m²), Heavy metal concentration (mg / kg). The calculation process integrates monitoring well data and soil pore water pressure sensor readings in real time, updating flux estimates every 30 minutes. The vertical cutoff wall depth control algorithm is based on the principle of limit equilibrium, considering the interaction of soil pressure, water pressure, and wall bending strength, outputting recommended insertion depth values (m) and construction parameters.

[0055] The hardware interface of the blocking device control submodule uses the industrial Ethernet protocol and is connected to the control system of the HDPE membrane insertion device. The depth adjustment instruction contains three operating parameters: target depth, insertion speed, and vibration frequency, which are transmitted to the hydraulic drive device through the field bus. The system is equipped with multiple safety interlocks that automatically pause construction when monitoring groundwater level changes or soil displacement abnormalities. The control effect feedback loop returns the permeability coefficient test results after the completion of the cutoff wall to the leakage flux calculation model, forming a closed-loop optimization mechanism.

[0056] The laser-induced breakdown spectroscopy system of the passivation agent reaction monitoring device uses a Nd:YAG pulse laser with an output wavelength of 1064 nm, a pulse energy of 100 mJ, and a repetition frequency of 10 Hz. The optical collection system consists of a lens group with a focal length of 75 mm and an optical fiber spectrometer, covering a spectral range of 200-900 nm. The sample chamber is designed as a flow-type reaction cell that receives the mixed liquid flow output by the grouting system in real time. The spectral data processing uses the partial least squares regression algorithm to establish a quantitative relationship model between the characteristic spectral line intensity of heavy metals and the chelation rate. The monitoring results are sent to the central control system through the wireless transmission module, with a data update interval of 5 minutes.

[0057] The parameter updating mechanism of the three-dimensional diffusion equation is based on forward difference approximation. When receiving the passivation agent effectiveness feedback signal, the retardation factor term in the equation is adjusted as follows: if the chelation rate increases by more than 10% of the baseline value, increase the retardation factor by 5%; if the chelation rate decreases by more than 15% of the baseline value, re-calibrate the adsorption kinetics parameters. The equation solver uses the algebraic multigrid method to accelerate convergence, and a full-field calculation takes about 8 minutes on a 16-core computing node. The updated parameter set is verified by sensitivity analysis and stored in the model knowledge base for optimization of subsequent prediction calculations.

[0058] The system operation monitoring interface integrates multiple functional views, including pollution plume three-dimensional visualization, real-time grouting parameter curves, and cutoff wall construction progress graphs. Operators can adjust control parameters through the touch screen, and important operations require double authentication. Historical data storage uses a time series database to record all sensor readings, control instructions, and system events, with a data retention period of 5 years. The remote diagnosis module allows technical support personnel to access the system operation status through a secure channel for fault diagnosis and parameter tuning.

[0059] The emergency response system contains three levels of processing plans: the first level of response is for local parameter exceeding, automatically adjusting the working parameters of a single grouting point; the second level of response is for regional abnormalities, coordinating multiple grouting points for joint adjustment; the third level of response is for system-level failures, starting the backup control unit and notifying the management personnel. All response actions record detailed logs, including trigger conditions, execution time, and operation results. The system regularly conducts simulation drills to test the processing flow and device status of each response level.

[0060] Example 4: Focus on the multi-dimensional data fusion and hierarchical response mechanism of the three-dimensional dynamic early warning platform. The actual operation case of a soil remediation project in an industrial park is used to illustrate the system workflow. The park covers an area of 86 hectares, and historical monitoring data shows that there is cadmium and lead composite pollution, with maximum concentrations of 28.7 mg / kg and 1560 mg / kg respectively. During the platform deployment stage, basic data collection is first completed, including core sample analysis of 12 drill holes, laboratory test results of 42 soil sampling points, and nearly five years of meteorological and hydrological observation records.

[0061] The core data processed by the pollution heat map generation engine comes from the spatial distribution map of heavy metals generated in Example 1, with the original data format being GeoTIFF files with a 50-meter grid. Data enhancement is implemented in the engine preprocessing stage, using the inverse distance weighting method to encrypt the grid to a resolution of 0.5 meters, considering the influence of soil type spatial variability during the interpolation process. The heat map rendering uses HSV color space conversion technology, defining five pollution levels corresponding to color bands: blue (<1 times standard value), green (1-2 times), yellow (2-3 times), orange (3-5 times), and red (>5 times). The three-dimensional visualization module is implemented based on WebGL technology, supporting interactive operations such as view rotation, profile cutting, and transparency adjustment, with VRAM memory usage controlled within 4GB.

[0062] The historical case library of the remediation scheme deduction module stores structured data of 327 projects, each case containing 14 types of feature parameters. The following table shows the comparison of key parameters of five typical remediation projects in the case library:

[0063]

[0064] The case matching algorithm uses an improved cosine similarity calculation method, defining a seven-dimensional feature vector: pollution type weight, concentration level, soil pH value, organic matter content, groundwater level, remediation target, and budget limit. After calculating the similarity between the current project feature vector and the case library, the top 10 most matched cases are output for the remediation scheme combination. The scheme optimizer is based on constraint satisfaction problem solving technology, considering the balance of time, cost, and technical feasibility, and finally generates three recommended paths: aggressive (focusing on effectiveness), balanced (considering both effectiveness and cost), and conservative (focusing on economy).

[0065] The comprehensive pollution risk index is calculated by integrating four dimensions: spatial distribution index (0-35 points), migration risk index (0-30 points), ecological sensitivity index (0-25 points), and exposure pathway index (0-10 points). The index update frequency is set to once an hour, and the calculation process uses a weighted summation model, with the weight coefficients determined by expert Delphi method. The risk level is divided into four levels: level I (<30 points) is acceptable risk, level II (30-50 points) is low risk, level III (50-70 points) is medium risk, and level IV (>70 points) is high risk.

[0066] The implementation of the multi-level risk response mechanism relies on the industrial Internet of Things platform, and the execution terminal includes PLC controllers, sound and light alarms, and access control systems. When the index reaches the primary warning threshold (70 points), the system automatically issues ModbusTCP instructions to the plant DCS system, reducing the pollution source area production load to 50% of the design value, and notifying the responsible engineer through the SMS platform. After the medium warning threshold (85 points) is triggered, the on-site warning light (red rotating) and buzzer (85 dB) are activated, and the emergency repair program is started simultaneously: the dosage of passivation agent is increased by 30%, the grouting frequency is changed to continuous mode, and the emergency team is put on standby. The highest warning threshold (95 points) response includes sending a site closure signal to the access controller, automatically locking all entrances and exits, turning on the emergency lighting system, and transferring control to the upper-level supervision platform through OPCUA protocol.

[0067] The human-computer interaction interface design of the three-dimensional dynamic early warning platform follows the standard, and the main console is equipped with a 55-inch 4K touch screen. The display area is divided into six functional blocks: the real-time monitoring area displays the three-dimensional heat map and risk index dashboard; the scheme deduction area presents the repair path comparison radar chart; the device status area displays the grouting system operating parameters; the alarm record area scrolls the abnormal events; the environmental parameter area updates meteorological and hydrological data; and the system log area records all operation traces. The interface supports multi-user collaborative operation, with different permission levels corresponding to differentiated function visibility, and key operations requiring iris identity authentication.

[0068] The data management subsystem uses a hybrid storage architecture, with real-time data stored in a time series database (sampling interval 1 minute), case library data stored in a relational database, and three-dimensional model data managed using a graph database. The data backup strategy sets daily incremental backup and weekly full backup, with off-site disaster recovery nodes set up in a backup data center 30 kilometers away. System security protection includes network layer firewall, application layer permission control, and data layer encrypted transmission, and passes the third-level information security certification.

[0069] The platform maintenance mechanism includes preventive maintenance and predictive maintenance. Preventive maintenance replaces key components according to fixed cycles: clean optical sensor lenses every 3 months, replace data center precision air conditioner filters every 6 months, and calibrate all monitoring probes annually. Predictive maintenance is based on equipment state monitoring data, predicts potential failures through LSTM neural networks, and generates maintenance work orders in advance. The system uses a dual-machine hot standby architecture, and the master and standby servers automatically switch through heartbeat detection, with a maximum fault recovery time of 90 seconds.

[0070] The industrial park implementation case shows that the system has cumulatively handled 12 risk warning events during 6 months of continuous operation, including 3 times of migration risk increase caused by heavy rain, 4 times of pollution increase caused by production abnormalities, and 5 times of repair interruption caused by equipment failure. The repair scheme generated by the platform is consistent with the actual engineering implementation results, and the median risk warning response delay time is 2 minutes and 48 seconds. During the system operation, a complete digital twin data chain is formed, providing a complete data basis for subsequent pollution site life cycle management.

[0071] Example 5: Around the collaborative operation mechanism of the intelligent grouting system and the passivation agent reaction monitoring device, the complete process flow from parameter analysis to efficiency feedback is described in detail. The hardware configuration of the intelligent grouting system adopts modular design, and the core components include Mitsubishi FX5U programmable controller, plunger type metering pump group, pressure buffer tank and distributed pipeline network. The controller receives the optimal passivation agent dosage signal from the pollution migration model through the Ethernet interface, which includes four parameters: target dosage concentration (g / m³), total grouting volume (m³), pressure upper limit (MPa) and flow range (L / min). The signal analysis module converts digital instructions into control variables, where the plunger pump stroke frequency is steplessly adjusted within 10-100 Hz according to the flow demand, and the pump group is designed in parallel to ensure that the maximum delivery capacity reaches 15 m³ / h.

[0072] The optical detection unit of the passivation agent reaction monitoring device adopts a closed flow path design, consisting of three functional parts: sample pretreatment cabin, laser action room and spectral analysis room. The sample pretreatment cabin implements online filtration, removes suspended particles through a 5 μm pore size ceramic filter, and the constant temperature system maintains the reaction liquid temperature at 25±0.5℃. The laser action room is equipped with beam shaping optics to control the laser spot diameter to 0.8mm, with a pulse energy stability better than ±2%. The spectral analysis room uses a echelle grating spectrometer system, matched with a back-illuminated CCD detector, to realize full-spectrum coverage of 200-900nm, with an optical resolution of 0.1nm. The device has an automatic calibration program, which performs wavelength calibration and intensity correction every 24 hours, and uses NIST standard reference materials to verify the system accuracy.

[0073] The dynamic adjustment process of grouting parameters is based on the feedforward-feedback composite control strategy. The feedforward control link calculates the initial working parameters of each grouting hole in advance according to the concentration gradient distribution of the heavy metal spatial distribution map. The feedback adjustment link collects the measurement values of the pipeline pressure sensor, electromagnetic flowmeter and online densimeter in real time, and adjusts the running state of the pump group through the PID algorithm. When the monitoring detects that the pipeline pressure fluctuation exceeds 15% of the set value, the pressure buffer tank automatically compensates the system fluctuation, and the buffer volume is designed to be 3 times the normal flow. The grouting terminal is provided with an intelligent distribution valve group, which dynamically adjusts the slurry distribution ratio of each grouting point according to the real-time update results of the pollution thermal map, and the distribution accuracy is controlled within ± 5%.

[0074] The data acquisition process of laser-induced breakdown spectroscopy technology includes four stages of pulse excitation, plasma formation, spectrum acquisition and data processing. 20 laser pulses are emitted in each analysis cycle, and the characteristic spectrum of the plasma cooled to 50 μs is collected through the gated detection technology. The heavy metal identification algorithm is based on spectral line database matching, focusing on monitoring the characteristic spectral lines of Cd 228.8 nm, Pb 405.8 nm and As 193.7 nm. The chelation rate calculation module compares the characteristic peak intensity ratio of free heavy metals and passivated heavy metals, and establishes a three-dimensional correction curve to eliminate the influence of matrix effect. Each batch of analysis results generates a data package containing time stamp, coordinate position and spectrum fingerprint, which is uploaded to the central database through industrial wireless network.

[0075] The transmission and processing of performance feedback signals adopt a layered architecture. The field layer device sends the raw monitoring data to the edge computing node through the IO-Link interface, and the node implements data cleaning and format standardization. The network layer uses Time-Sensitive Network (TSN) protocol to ensure the real-time performance of data transmission, and the worst-case end-to-end delay is less than 50 ms. The platform layer deploys a stream processing engine to analyze the continuously arriving performance data in a sliding window, with a window size of 30 minutes of recent data, and calculates the moving average as the current performance indicator. When it is detected that the chelation rate has decreased by more than 10% for three consecutive windows, the system generates a parameter update request and pushes it to the pollution migration model.

[0076] The cooperative working mechanism of the grouting system and the monitoring device is reflected in three key interaction links. First, the parameter pre-setting before grouting starts, according to the historical data of the monitoring device to predict the best laser detection parameters, to avoid too high energy causing spectral saturation or too low energy causing insufficient signal-to-noise ratio. Second, adaptive adjustment during operation, when the chelation efficiency of a certain area is abnormal, automatically increase the sampling frequency of the grouting point in that area to 3 times the normal value. Third, the cooperative arrangement of the maintenance period, combined with the planned downtime of the grouting system to carry out preventive maintenance of optical devices, to minimize system downtime.

[0077] Emergency handling mechanism sets two-level response process. Primary response targets local performance decline events, automatically triggers flushing procedure of the grouting point, uses pH = 5 dilute nitric acid solution to clean the pipeline and mixer, duration 2 minutes. Advanced response handles system-level failures, such as laser energy attenuation exceeding 40% or spectrometer CCD temperature exceeding 50℃, immediately starts the backup analysis module and notifies the maintenance personnel. All emergency events record detailed logs, including trigger time, disposal measures and recovery status, log files use WORM storage technology to prevent tampering.

[0078] System calibration and verification system contains four levels. Device-level calibration performs optical system energy test and wavelength calibration every day, uses neodymium glass standard sample to verify laser performance. Method-level calibration performs standard sample test every week, covering 5 concentration gradients of heavy metal solution. System-level verification is implemented once a month, by injecting known concentration of standard material through the grouting system, tracking the analysis detection accuracy throughout the process. External comparison is carried out once every quarter, correlating online monitoring results with laboratory ICP-MS analysis data, maintaining a consistency standard of R²>0.95.

[0079] Human-computer interaction interface design focuses on operation intuitiveness and information integrity. The main control screen is divided into real-time monitoring area, historical trend area, alarm information area and operation control area. Real-time monitoring area dynamically displays grouting pressure curve, flow distribution graph and chelation rate thermodynamic diagram, with an update frequency of 1 Hz. Historical trend area can retrieve parameter change curves of any time period, supporting multivariate superimposed display. Alarm information area displays current alarms according to severity classification, with disposal suggestion guidance. Operation control area implements hierarchical management of permissions, routine parameter adjustment is authorized to site operators, key parameter modification requires engineer-level permission. Interface color scheme uses blue-gray as the main color in line with human factors engineering, important alarms use orange flashing prompt.

[0080] Data management subsystem realizes full life cycle tracking. Raw spectral data retains complete waveform information, uses HDF5 format for compressed storage, a single file contains 1000 pulse acquisition data. Process parameter data is stored in time series database, label system contains device ID, parameter type and timestamp triple index. All data save period is set to 5 years after the end of the project, supports multi-dimensional retrieval by time range, spatial location and heavy metal species. Data export function generates report files in line with EPA standards, including detection results, quality control data and meta-information description.

[0081] The mechanical design of the intelligent grouting system takes into account the adaptability to harsh environments. The pump set and pipeline are made of 316L stainless steel, with corrosion resistance meeting the working range of pH 2-11. The laser optical assembly is installed in a shockproof constant-temperature cabin, with the environmental temperature controlled between 20-30°C and the relative humidity maintained at 30%-60%. The electrical system reaches an IP65 protection level, with redundant backup for key lines. The layout of field devices follows the principle of modularity, with each functional unit maintaining a maintenance channel of more than 1.2 meters, and heavy components equipped with lifting rings for easy maintenance and replacement.

Claims

1. A soil heavy metal pollution identification system, characterized in that, The application relates to a pollution risk assessment system, which comprises the following modules: a multispectral remote sensing perception module, which is configured with a high-resolution imaging unit and a ground spectrum acquisition unit, the high-resolution imaging unit acquires ground reflectivity data through a satellite load, the ground spectrum acquisition unit collects in-situ soil spectrum data by using a vehicle-mounted mobile platform, the multispectral remote sensing perception module is integrated with a heterogeneous data fusion gateway, the heterogeneous data fusion gateway receives the ground reflectivity data and the in-situ soil spectrum data and generates a multi-band spectral response signal; a pollution risk assessment module, which comprises a spatial distribution analysis unit, a migration risk prediction unit and a pollution threshold definition unit, the spatial distribution analysis unit processes the multi-band spectral response signal and outputs a heavy metal spatial distribution atlas, the migration risk prediction unit generates a pollution migration probability cloud map in combination with the heavy metal spatial distribution atlas and meteorological and hydrological data, and the pollution threshold definition unit defines ecological safety constraints according to the multi-band spectral response signal and outputs a soil remediation safety threshold; a governance decision execution module, which comprises an in-situ remediation execution unit, a pollution source control unit and a three-dimensional dynamic early warning platform, the in-situ remediation execution unit receives the pollution migration probability cloud map and adjusts the injection parameters of a passivation agent, the pollution source control unit regulates the operation state of a pollution source blocking device based on the soil remediation safety threshold and the multi-band spectral response signal, and simultaneously collects soil pore water real-time monitoring signals, and the three-dimensional dynamic early warning platform integrates the heavy metal spatial distribution atlas and the pollution migration probability cloud map to generate a comprehensive pollution risk index.

2. The soil heavy metal pollution identification system of claim 1, wherein The multispectral remote sensing perception module further comprises a UAV near-ground detection unit and a groundwater quality sensing network, the UAV near-ground detection unit obtains vegetation stress characteristic data by carrying a hyperspectral sensor on a multi-rotor flight platform, the groundwater quality sensing network is arranged in a pollution site monitoring well and collects underground water ion concentration data, the heterogeneous data fusion gateway receives the vegetation stress characteristic data and the underground water ion concentration data to perform space-time alignment processing, forms a multi-band spectral response signal containing spectral characteristic signals, vegetation response signals and ion diffusion signals, and the multi-band spectral response signal is specifically divided into spatial analysis spectral signals, migration prediction spectral signals and threshold definition spectral signals.

3. The soil heavy metal pollution identification system of claim 2, wherein The spatial distribution analysis unit adopts a convolutional neural network to construct a geostatistical inversion model, the spatial analysis spectral signals contain visible-near-infrared band reflectivity data, short-wave infrared absorption characteristic data and historical pollution census data, the geostatistical inversion model extracts spectral spatial features through a depth separable convolutional layer, inputs the visible-near-infrared band reflectivity data, the short-wave infrared absorption characteristic data and the historical pollution census data into a three-dimensional convolution kernel to generate a heavy metal spatial distribution atlas, and simultaneously outputs pollution source tracing results by fusing the vegetation response signals and the ion diffusion signals through a feature pyramid network.

4. The soil heavy metal pollution identification system of claim 3, wherein The migration risk prediction unit includes a pollution migration model and a Monte Carlo simulator, the migration prediction spectrum signal includes soil texture parameters, groundwater flow rate data and rainfall erosion intensity data, the pollution migration model processes the migration prediction spectrum signal by coupling the convection-diffusion equation with the adsorption-desorption kinetics equation, outputs the heavy metal migration flux prediction value, the Monte Carlo simulator generates a pollution migration probability cloud based on the heavy metal migration flux prediction value through 1000 times of random sampling operation, and the pollution migration probability cloud is fed back to the geo-statistical inversion model to update the pollution source tracing result.

5. The soil heavy metal pollution identification system of claim 2, wherein The pollution threshold defining unit processes the threshold defining spectrum signal through an ecological toxicology constraint model, the ecological toxicology constraint model is integrated with a species sensitivity distribution analysis module, the species sensitivity distribution analysis module calls biological half lethal concentration data in a local biological database to define ecological safety constraint conditions, outputs a soil remediation safety threshold, and simultaneously dynamically adjusts the soil remediation safety threshold to be not lower than a legal limit standard by using a model predictive control algorithm.

6. The soil heavy metal pollution identification system of claim 5, wherein The in-situ remediation execution unit is configured with a pollution migration model and a passivation agent optimization library, the pollution migration model constructs a three-dimensional diffusion equation based on the second law of Fick, input parameters of the three-dimensional diffusion equation include soil oxidation-reduction potential data, organic matter content data and the pollution migration probability cloud, and output parameters include an optimal passivation agent dosage signal, the passivation agent optimization library stores spectral response characteristic data of historical remediation materials, and a passivation agent type selection instruction is generated by matching the spectral response characteristic data with the optimal passivation agent dosage signal.

7. The soil heavy metal pollution identification system of claim 6, wherein The pollution source control unit includes a pollution flux monitoring sub-module and a blocking equipment control sub-module, the pollution flux monitoring sub-module receives real-time soil pore water monitoring signals and heavy metal spatial distribution maps, calculates heavy metal leakage flux in a pollution source area, and the blocking equipment control sub-module generates a vertical barrier wall depth adjustment instruction based on the difference between the heavy metal leakage flux and the soil remediation safety threshold, the vertical barrier wall depth adjustment instruction is transmitted to the three-dimensional diffusion equation to update the optimal passivation agent dosage signal.

8. The soil heavy metal pollution identification system of claim 1, wherein The three-dimensional dynamic early warning platform is deployed with a pollution thermal map generation engine and a remediation scheme deduction module, the pollution thermal map generation engine renders the heavy metal spatial distribution map into a three-dimensional pollution thermal map by using a Kriging interpolation algorithm, the remediation scheme deduction module generates a multi-scheme remediation path based on the comprehensive pollution risk index by calling a historical remediation case library, and the multi-scheme remediation path is input into the in-situ remediation execution unit to trigger dynamic adjustment of passivation agent parameters.

9. The soil heavy metal pollution identification system of claim 8, wherein The three-dimensional dynamic early warning platform implements a multi-level risk response mechanism, when the comprehensive pollution risk index reaches a primary warning threshold, a pollution source area production reduction instruction is started, when the comprehensive pollution risk index reaches an intermediate warning threshold, a pollution site isolation protection instruction is activated and the in-situ remediation execution unit is triggered to urgently add passivation agents, and when the comprehensive pollution risk index reaches a highest warning threshold, a site emergency closure signal is sent to the governance decision execution module.

10. The soil heavy metal pollution identification system of claim 6, wherein The in-situ remediation execution unit specifically comprises an intelligent grouting system and a passivation agent reaction monitoring device, the intelligent grouting system analyzes the optimal passivation agent dosage signal and dynamically adjusts the grouting pressure and the grouting flow, the passivation agent reaction monitoring device detects the heavy metal chelation rate of the passivation agent reaction product in real time through a laser-induced breakdown spectroscopy technology, generates a passivation agent performance feedback signal, and the passivation agent performance feedback signal is transmitted to the pollution migration model to update the three-dimensional diffusion equation parameters.

Images